CAISO 15-Min Liquidity

Transcription

1 CAISO 15-Min Liquidity CAISO Stakeholder Meeting October 6, 2015 Alex Spain Trading Floor Manager Power Services BPA Dave Dernovsek Day Ahead Trader Power Services BPA

2 Presentation Objectives Federal Columbia River Power System (FCRPS) Facts and Myths CAISO HASP/FMM Liquidity Impediments Operational Considerations Transmission Utilization Market Issues CAISO HASP/FMM Liquidity Solutions 2

3 FCRPS: Facts FCRPS is a partnership US Army Corps of Engineers (Corps) US Bureau of Reclamation (Reclamation) Bonneville Power Administration (BPA) FCRPS 31 hydro plants 21 Corps (14,651 MWs) 10 Reclamation (7,807 MWs) 209 Generating Units ranging from 1 MW (Boise Diversion) to 805 MW (Grand Coulee) Canada has 15% basin area, but provides 30% of ave annual flow at The Dalles 3

4 FCRPS: Facts Multiple FCRPS Objectives Flood Control Irrigation Navigation Recreation Fish Operations (BiOps) Control Area Services Power 4

5 FCRPS Myth #1 FCRPS has lots of flexibility Average January - July runoff is 106 million acre feet (ranges from MAF) Federal storage about 30 MAF: Storage Limited System When the FCRPS is empty, we can store 25 % of the annual runoff. The Colorado or Missouri systems can store two to three times the annual runoff. 12 Hydro Projects provide 94% of gen capacity Generation is driven by need to move water. ~1000 MWs of INC/DEC Capacity is allocated to Gen/Load Imbalance FCRPS persistently faced Spring Oversupply due to lack of flexibility 5

6 FCRPS Myth #2 FCRPS is always Surplus White Book (WBK 2014) 1958 (Ave Water Year) and 1937 (Critical Low Water Year) Graph shows % of FCRPS capacity and energy allocated to LT Obligations Balanced Water Year Driven Oversupply Risk 6

7 FCRPS Myth #3 Trading Floor controls majority of FCRPS Capacity ~ 90% of BPA Power Services revenues and FCRPS Flexibility is committed prior to Trading Floor transactions 7

8 BPA Trading Floor Market Transactions support BPA needs to reshape natural flows to meet operational requirements prior to Real Time Typical Annual Operations Cycle 500, , , , , , , , ,000 50,000 O N D J F M A M J J A S 0 Natural Flow TDA Flow

9 BPA Trading Floor Goal is to achieve load service and operational requirements prior to real time (Hourly)

10 BPA Trading Floor CA/CAISO volume is capped by intertie transmission capacity 143 MAF 129 MAF 98 MAF 108 MAF 85 MAF

11 BPA Trading Floor HASP/FMM Price and Dispatch Uncertainty introduces costs and risks Risky

12 BPA Trading Floor CAISO DA avoids HASP/FMM price and dispatch costs and risks

13 NOB/Sylmar (FMM HASP) Spread CAISO DA bids hedge recurring HASP/FMM price volatility Recurrent periods of volatile FMM/HASP spread

14 FMM-HASP Spread Frequency and Bias HASP < FMM HASP > FMM HASP < FMM HASP > FMM Heavy tail on Sylmar where HASP price is more than $100 below eventual FMM

15 NOB/Sylmar: Implications of FMM-HASP Spread Bias Risk of ineffective economic hourly bid strategy incentivizes resources to self-schedule as a price taker or in the DAM Sylmar hourly block bidding - inc Unit opportunity cost: $ - 5/1/14-8/31/15 Transmission cost: $ 2.00 bid price MWh HASP $/MWh FMM $/MWh revenue p/l price taker 11,875 $ $ $ 396,363 $ 372,613 $ - 10,924 $ $ $ 383,572 $ 361,724 $ ,660 $ $ $ 375,168 $ 353,848 $ ,454 $ $ $ 370,596 $ 349,688 $ ,175 $ $ $ 363,700 $ 343,350 $ ,594 $ $ $ 350,591 $ 331,403 $ ,054 $ $ $ 312,766 $ 296,658 $ ,819 $ $ $ 249,916 $ 238,278 $ ,809 $ $ $ 182,857 $ 175,239 $ ,083 $ $ $ 116,607 $ 112,441

16 NOB/Sylmar: FMM-HASP Spread Bias Costs FMM HASP Spread risk increases when CAISO sends a HASP signal to incentivize Exports.

17 NOB/Sylmar: Cost Implications of FMM/HASP Bias Positive HASP signal does not protect against negative binding FMM price in almost 3% of hours Sylmar FMM dec prices 5/1/14-8/31/15 # intervals avg HASP avg FMM FMM - HASP All instances when avg FMM < $ (29.24) $ (38.20) $ (8.96) instances where HASP > 0 and FMM < $ $ (27.94) $ (53.92)

18 COB/Malin (FMM HASP) Spread Greater price convergence but still evidence of upward bias

19 COB/Malin: FMM Price Protection with Stranded Costs Stranded firm intertie and FCRPS generation capacity more prevalent in FMM compared to DAM or Hourly Block CAISO Market Transactions - ratio of flow to awarded MWh 12/1/2014-8/31/2015 COB+NOB COB only DAM 99.47% 99.05% Hourly Block 98.97% 98.58% FMM 75.67%

20 COB/Malin: FMM Price Protection with Stranded Costs In absence of alternative compensation, FMM volatility alone may be insufficient to recover the costs associated with providing CAISO 15-min dispatch option Decreasing FMM interval volatility

21 HASP Liquidity impediments for BPA FCRPS operational requirements generally met in preschedule There are structural challenges to participating in CAISO economic hourly block markets, specifically persistent FMM/HASP spread bias Main cause of FMM/HASP divergence traced to HASP clearing near or below price floor of -$150 High tail of positive FMM/HASP divergence affects overall strategy when attempting to utilize CAISO economic real-time markets Economic bidding at low opportunity costs in Hourly Block becomes suboptimal compared to self-scheduling as a price taker There will be no liquidity in dec market if market participants cannot rely on HASP to send a realistic price signal

22 FMM Liquidity impediments for BPA FMM Market provides no Operational Benefits Real-Time markets small portion of overall transaction portfolio FCRPS operational requirements generally met on an hourly basis BPA has limited incentive to expand FMM participation Need economic incentive to lose a quarter of transmission while giving free option to CAISO Interval FMM volatility may not always be sufficient to compensate intertie and flexible generator owners for the option value

23 HASP and FMM Potential Liquidity Solutions Greater FMM/HASP price convergence at nodes where FMM is not available Create greater liquidity in both import and export economic bids Potential to reallocate transactions from DA to RT Compensate for reduced flow and resource utilization in FMM Capacity/FlexiRamp/Mileage charge Interval FMM volatility sufficient to incentivize FMM over Hourly Block Introduce 15-minute scheduling on the DC More detailed discussion of contributing factors to market price signals Inelastic supply stack (marginal MW that binds the tie is a self-schedule) Disparity of 15-minute scheduling ability across TSPs Seams issues: scheduling practices, intertie TTC assumptions Load/generation variability and changes in modeling assumptions (Transmission Reliability Margin, Operator adjustments)

24 Questions QUESTIONS?